Uranium dioxide is used to produce ceramic UO.sub.2 pellets for nuclear power plants, or as a source for production of uranium tetrafluoride, UF.sub.4. The UF.sub.4 is then further used to produce metallic uranium, or is converted to uranium hexafluoride, UF.sub.6, by fluorination. Conventional methods for producing UO.sub.2 typically yield a product with an oxygen to uranium ratio of 2.04 or greater. It is important that the uranium dioxide feedstock for fluorination have an oxygen to uranium ratio close to 2.00. Uranium dioxide feedstocks in which this ratio is significantly higher than 2.00 cause formation of uranyl fluoride, UO.sub.2 F.sub.2, which contaminates the UF.sub.4 product. This reduces yields and increases the cost of fluorination.
One method for preparing uranium dioxide is to reduce pulverized uranium trioxide powder in a hydrogen atmosphere, usually in a conventionally heated fluidized bed, at a temperature of about 7000.degree. C. The chemical conversion is represented by equation (1) EQU UO.sub.3 +H.sub.2.fwdarw.UO.sub.2 +H.sub.2 O (1).
The UO.sub.2 obtained under these conditions normally has a minimum O:U ratio of 2.04. To reduce this ratio further, two approaches are known: 1) the particle size of the UO.sub.3 starting material is reduced to the micrometer range; or 2) a higher temperature is used. Neither of these solutions is particularly satisfactory. To obtain a sufficiently small UO.sub.3 particle size, special precipitation techniques may be required. Such techniques are time-consuming, often difficult to carry out, and expensive. Option 2), increasing the temperature of a fluidized bed reactor, requires a considerable increase in energy consumption. Heat transfer in such reactors is highly inefficient due to low thermal conductivity of the bed, and significant heat is lost with the flow of fluidizing gas.
The fluidized bed method suffers from other drawbacks. High flow rates of carrier gas and hydrogen are required to support the fluidization, with the result that hydrogen consumption is approximately 170% to 190% of the stoichiometric amount indicated by equation (1). The UO.sub.3 starting material must be pulverized prior to feeding into the reactor. The process is highly sensitive to particle size distribution. Furthermore, high maintenance costs are incurred because heating elements must be periodically replaced, requiring shutting down of the reactor.
It has been proposed to produce sinterable UO.sub.2 by the use of microwave heating [Canadian patent No. 1,197,069; Thornton, Thomas A.; Holaday, Veldon D., Jr., which is incorporated herein by reference]. Thornton et al propose a process that involves the absorption of microwave radiation by uranyl nitrate hexahydrate (UNH), ammonium diuranate (ADU) or ammonium uranyl carbonate (AUC). These uranium salts are decomposed, preferably in an oxidizing atmosphere, at elevated temperatures, to yield an intermediate product which may have a uranium oxide stoichiometric range of from UO.sub.3 to U.sub.3 O.sub.8. The intermediate product is further heated in a microwave furnace, in a reducing atmosphere, to reduce it to sinterable uranium dioxide powder.
Thornton et al, in U.S. Pat. No. 4,389,355, which is incorporated herein by reference, have proposed a process for preparing nuclear fuel pellets that involves sintering UO.sub.2 powder and an organic binder in a microwave induction furnace in a reducing atmosphere. Sintered compacts are cooled under reducing atmospheric conditions and then ground to the desired finished uranium dioxide pellet product. Scrap uranium dioxide powder and rejected pellets are recycled to a microwave induction furnace where they are heated in an oxidizing atmosphere to convert UO.sub.2 to U.sub.3 O.sub.8. The U.sub.3 O.sub.8 is then blended with UO.sub.2 and organic binder at the beginning step of the nuclear fuel pellet preparation process.
Ford and Pei, in the Journal of Microwave Power, 2--2, 1967, pages 61 to 64, the disclosure of which is incorporated herein by reference, propose heating various materials, including uranium dioxide, by microwave radiation.
Paul Haas discusses heating uranium oxides in a microwave oven; see page 873 of the American Ceramic Society Bulletin, Volume 58, No. 9, (1979) the disclosure of which is incorporated herein by reference. Haas found that UO.sub.2 and U.sub.3 O.sub.8 samples heated strongly under microwave irradiation. Dry UO.sub.3 samples did not show any significant heating when exposed in a microwave oven. However, in tests on microwave drying of samples of hydrated UO.sub.3 gel spheres, hot spots were observed to develop. Haas suggested that the hydrated gel first underwent small amounts of reduction from traces of NH.sub.3 and organic materials in the gel. Once overheating started, hexavalent uranium was converted to U.sub.3 O.sub.8, which absorbs microwave energy.
Use of microwave energy to heat oxides of uranium Is also discussed by Van Loock and Tollenaere, Sprechsaal, Vol 122, No. 12, 1989, pages 1157-1159 and by Sturcken and McCurry, in Ceramic Transactions 1991, Volume 21, pages 117-123, the disclosures of which are incorporated herein by reference.
A problem encountered when microwave energy is used to heat uranium oxides to convert them to UO.sub.2 is that the microwave energy is absorbed and attenuated in the outer layer of the irradiated uranium oxides, where the UO.sub.2 product is first formed, and does not penetrate deeply into the material. The interior of the material is screened by the outer layer of UO.sub.2, so that the microwave energy is absorbed in the outer layer and does not penetrate to the interior. This leads to overheating of the outer layer (thermal runaway), resulting in non-uniform heating, and a cool reactor core. Ford and Pei encountered non-uniform heating of uranium dioxide which, they say, was very unsatisfactory and caused a suspension of experiments. It may be possible to reduce this screening effect and to avoid non-uniform heating by sophisticated powder mixing techniques, but these techniques are not practical on a large industrial scale.
Thus, the idea of direct microwave heating of uranium oxides with microwaves in a single mode resonant cavity or multimode oven applicators has been demonstrated with small loads, less than about 50 grams, by the several above-mentioned workers, whose publications are incorporated herein by reference. However, this technique is not feasible for larger scale applications, where the dimensions of processing material load are much greater, so that the desired length of the path of the radiation in the material is much greater than the penetration depth of the microwave radiation. The penetration depth of microwave radiation into UO.sub.2 is in the millimeter or centimeter range. Due to the attenuation of the microwave power in the outer layers of the reduced UO.sub.2 product, the inner part of the processing material load will not be irradiated, and heating of the inert part occurs by thermal conductivity and load mixing only.